Cooling The City – Green Facades For Mitigating Urban Heat

Anyone who’s ever visited an Indian city in the heat of summer will remember the feeling that the city itself is baking – heat radiates from the pavements, the buildings and there are fewer green areas to mitigate these effects. This is the urban heat island effect – when the city is much hotter than the surrounding areas.

Urban heat islands are a common problem and are only going to get worse as the climate warms. This is especially true in countries closer to the equator – summer is hot but is now promising to get even hotter.

One of the solutions that people have been looking at is “greening building facades” or more simply – growing plants along the walls of buildings – vertical greening. The advantages of “green building facades” are that the plants help reduce building temperatures through shading by the leaves, reducing the impact of wind and through evapotranspiration. Now, like any other natural system – growing plants vertically means that there are a number of parameters that have to be controlled or monitored to get the desired effect. Things like humidity, soil characteristics, the choice of plants….

While there are quite a few studies that looked at what happens for a single building, it’s more useful for city planners or companies with large campuses to figure out how to make vertical greening effective at a larger scale. This is where a study in Berlin could provide useful hints. The authors used a Bayesian network to model the impact of different policies on vertical greening. In their model, “Actors, planned interventions, and unchangeable parameters are displayed as Elements in a graph.

Conditions of certain elements affect the state of other elements, which is shown by directed links between the elements. The probabilities of various possible states are calculated and the conditional probability for the final outcome - i.e. the implementation of a green facade is determined".

The result from their study showed that their are a number of factors (attitudes, financial inducements, actual performance etc) that influence building a green facade through out a city and that these factors can be ranked. In Berlin, the most important factors they found were "positive attitudes towards green facades" and financial incentives.

Where something like this helps is that it helps people identify what would work best locally (either at a city scale or a neighborhood or campus scale) and the factors that are needed to make it happen.  From a technical standpoint, what’s interesting about what they did, apart from building a complex Bayesian model, is that they used feedback from experts to better calibrate their model. This is a classic example of refining a fairly traditional data science algorithm (Bayesian models) with expertise in the subject area - again something that people who’re interested in the intersection of the two fields would find valuable.